A pantograph operation state monitoring analysis method and system

Abstract

The application discloses a pantograph operation state monitoring analysis method and system, relates to the field of track train monitoring, and mainly has the following scheme: through comprehensive analysis on vibration data, contact force data, contact point temperature data and abrasion data, whether a pantograph pressure control instruction is triggered, whether a train speed limit instruction is triggered and whether a pantograph operation failure exists are judged, a passive early warning strategy is changed into an active regulation and control, the operation state of the pantograph can be more comprehensively evaluated, through intelligent monitoring and prediction technology, the frequency and cost of manual detection are reduced, the evaluation accuracy is improved, and meanwhile, maintenance cost and safety risks caused by failure expansion are avoided.

Description

Technical Field

[0001] This invention relates to the field of rail train monitoring technology, specifically to a method and system for monitoring and analyzing the operating status of a pantograph. Background Technology

[0002] In rail transit systems, the pantograph plays a crucial role in obtaining electrical energy from the overhead contact line and transmitting it to the train. The stability of its operation directly affects the reliable operation of the train and the safety and operational efficiency of the entire rail transit system. Therefore, developing an effective method for monitoring and analyzing the pantograph's operating status is of great significance for ensuring train operation safety and improving transportation efficiency.

[0003] Currently, the monitoring of pantograph operation status mainly relies on traditional detection technologies. For example, Chinese patent document CN204495300U discloses a pantograph operation status monitoring device. By collecting image data information of the pantograph, the device can obtain image data information of the pantograph and detect the pantograph size information, such as the thickness of the pantograph slide plate, in a timely, convenient and fast manner, thereby ensuring that the pantograph is in normal operating condition.

[0004] While existing technologies can reflect the operating status of pantographs to some extent, they have the following drawbacks: 1. They rely on only a single type of data, resulting in a limited data evaluation model and an inability to comprehensively reflect the pantograph's operating status. 2. They employ simple statistical analysis or threshold judgments, lacking in-depth data mining and comprehensive analysis, leading to limited accuracy and reliability of diagnostic results. 3. Traditional methods suffer from time errors in multi-data processing and analysis, making it difficult to accurately assess the pantograph's operating status. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] To address the shortcomings of existing technologies, this invention provides a method for monitoring and analyzing the operating status of pantographs, which at least solves the problem that existing technologies have a single type of reference parameter and insufficient accuracy in assessing the operating status of pantographs.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring and analyzing the operating status of a pantograph, comprising:

[0009] Step 1: Synchronously collect vibration data, contact force data, contact point temperature data, and wear data of the pantograph using a multimodal sensor system;

[0010] Step 2: Perform time alignment and data preprocessing on the collected data;

[0011] Step 3: By comprehensively calculating the processed vibration data, contact force data, and contact point temperature data, the dynamic contact force of the pantograph is generated, and the pantograph pressure control command is triggered based on the dynamic contact force.

[0012] Step 4: Analyze the vibration data and dynamic contact force, and calculate the probability of hard point impact; determine whether to trigger the train speed limit command based on the probability of hard point impact.

[0013] Step 5: By comprehensively analyzing wear data, contact point temperature data, and contact force data, the predicted damage depth is obtained; the wear area value of the pantograph friction area is collected through a quantum dot fluorescence monitoring system; and the dynamic weights of dynamic contact force, hard point impact probability, predicted wear depth, and wear area value are calculated using an adaptive model.

[0014] Step 6: Extract features from the preprocessed dynamic contact force, hard point impact probability, predicted wear depth, and wear area value respectively. Then, input the extracted different features and corresponding dynamic weights into the multimodal data fusion neural network to output the multimodal fusion feature vector.

[0015] Step 7: Input the multimodal fusion feature vector into the operation status assessment model, and use the operation status assessment model to assess whether there is an operational fault in the pantograph.

[0016] In the preferred scheme of the pantograph operation status monitoring and analysis method, the method for time alignment of the collected data is as follows:

[0017] A unified timescale is added to each sensor in the multimodal sensor system, and then the data transmission time difference of each sensor is calculated using an optical fiber transmission delay compensation algorithm. The compensation formula is as follows:

[0018] ;

[0019] Among them, t 补偿 The time difference that needs to be compensated for during data transmission from the sensor; L is the fiber optic length; n 纤 c is the refractive index of the fiber core; t is the speed of light; 处理 The processing delay time of the demodulator during the data transmission process of the sensor;

[0020] Calculate t for each sensor 补偿 Select a reference time point t 参考 Then calculate the time delay for each sensor. t, time delay The formula for calculating t is: t=t 参考 -t 补偿 ;

[0021] Based on the time delay of each sensor t adjusts the timestamp of the corresponding sensor to match the reference time point t. 参考 Alignment.

[0022] In the preferred scheme of the pantograph operation status monitoring and analysis method, the method for generating the dynamic contact force of the pantograph is as follows:

[0023] The preprocessed vibration data, contact force data, and contact point temperature data are input into the data coupling model for comprehensive calculation, outputting the dynamic contact force of the pantograph. The calculation formula of the data coupling model is as follows:

[0024] ;

[0025] in, F d It is the dynamic contact force of the pantograph; F s It is the static contact force of the pantograph; m i It is the vibrating mass of the pantograph in different directions. i The index indicates different directions and can be 1, 2, or 3. m 1 represents the component of the pantograph's vibrating mass in the X-axis direction; m 2 represents the component of the pantograph's vibrating mass in the Y-axis direction; m 3 represents the component of the pantograph's vibrating mass in the Z-axis direction; a 1 represents the vibration acceleration component of the pantograph in the X-axis direction; a 2 represents the vibration acceleration component of the pantograph in the Y-axis direction; a 3 represents the vibration acceleration component of the pantograph in the Z-axis direction; β ρ is the coefficient of thermal expansion; C is the density of air. d A is the aerodynamic drag coefficient; A is the frontal area of ​​the vehicle body. T 0 is the reference temperature; T c It is the temperature of the pantograph's contact point; v It's the vehicle speed.

[0026] In the preferred scheme of the pantograph operation status monitoring and analysis method, the method for determining whether a pantograph pressure control command has been triggered is as follows:

[0027] Set the contact force range threshold;

[0028] The system collects dynamic contact forces within a unit time period to form a dynamic contact force sequence. It calculates the average contact force of multiple dynamic contact forces in the sequence and compares the average contact force with a contact force interval threshold. If the average contact force does not meet the contact force interval threshold, a pressure control command needs to be triggered.

[0029] In the preferred scheme of pantograph operation status monitoring and analysis method, the method for calculating the probability of hard point impact by analyzing vibration data and dynamic contact force is as follows:

[0030] Vibration data and dynamic contact force are input into the hard-point impact probability model. The hard-point impact probability model can extract the kurtosis coefficient from the vibration data, as shown in the formula: Where K is the kurtosis coefficient. N σ is the number of input samples; μ is the standard deviation of the pantograph's acceleration; and μ is the mean of the pantograph's vibration acceleration components.

[0031] The hard-point impact probability model can extract the contact force fluctuation rate from the dynamic contact force sequence, as shown in the formula: ;in, δ F For contact force fluctuation rate, F d,max and F d,min These are the maximum and minimum values ​​of the dynamic contact force in the dynamic contact force sequence, respectively. F d,avg It is the mean of the dynamic contact forces in the dynamic contact force sequence; then, the hard point impact probability is calculated and output using the determination formula, which is:

[0032] ;

[0033] The method for determining whether a train speed limit command is triggered based on the probability of hard point impact is as follows:

[0034] Set hard point impact assessment thresholds;

[0035] When the probability of a hard impact is greater than or equal to the hard impact assessment threshold, a train speed limit command is triggered.

[0036] In the preferred scheme of the pantograph operation status monitoring and analysis method, the method for predicting damage depth is as follows:

[0037] The network structure of the wear depth prediction model includes an input layer, an LSTM layer, and an output layer. The input layer...

[0038] Wear data, contact point temperature data, and contact force data from a past unit of time period are input into the input layer of the wear depth prediction model;

[0039] The analysis and calculation are performed using an LSTM layer, and the formula is as follows:

[0040] ,in, d s It is the predicted wear depth of the pantograph. dw It is the wear depth of the pantograph, t 当前 It is the current time point, t 单 It is a unit of time period;

[0041] The output layer outputs the predicted wear depth for a future unit of time period.

[0042] In the preferred scheme of the pantograph operation status monitoring and analysis method, the method of calculating dynamic weights through an adaptive model is as follows:

[0043] The dynamic contact force, hard point impact probability, predicted wear depth, and wear area values ​​from historical time periods are integrated into a multimodal dataset. This multimodal dataset, along with vehicle speed, wind speed, and contact point temperature at corresponding time points, is then input into an adaptive model to calculate the dynamic weight of each parameter in the multimodal dataset, as shown in the following formula:

[0044] ;

[0045] Where, q z is the dynamic weight of different data in the multimodal dataset, and z represents the sequence number of different data in the multimodal dataset. The sequence number can be 1, 2, 3 or 4, which correspond to the sequence number of dynamic contact force, hard point impact probability, predicted wear depth and wear area value, respectively. S z This indicates the parameter sensitivity of different data in a multimodal dataset; E z represents the information entropy of different data in a multimodal dataset.

[0046] In the preferred scheme of the pantograph operation status monitoring and analysis method, the extracted different features and corresponding dynamic weights are input into a multimodal data fusion neural network, and the multimodal fusion feature vector is output. ,in, It is the wear area value of the wear zone.

[0047] In the preferred scheme of pantograph operation status monitoring and analysis method, the method for predicting whether there is a risk of operational failure in the pantograph within a future unit of operating time is as follows:

[0048] Several training samples are constructed using multimodal fusion feature vectors Rr from historical records and their corresponding fault category labels Gl, forming a training sample set. ,in, Rr n Indicates the first n The multimodal fusion feature vector of each training sample; Gl n Indicates the first n Fault category labels for each training sample; nThis is the index of the training sample in the training sample set, and its value is a positive integer;

[0049] The support vector machine model is trained using a training sample set. After training, the multimodal fusion feature vector Rr of new samples is input into the support vector machine model to calculate the classification function of the new samples. f(x) By classification function f （x） The numerical values ​​are used to match the fault type of new samples; classification function f(x) The formula is:

[0050] ;

[0051] Where, ax j For Lagrange multipliers; Rr j Let Gl be the multimodal fusion feature vector of the j-th sample. j Let j be the fault category label for the j-th sample. b is the kernel function; b is the bias term.

[0052] The calculation formula is:

[0053] ;

[0054] Where γ is the kernel function parameter, which controls the width of the kernel function.

[0055] (III) Beneficial Effects

[0056] This invention provides a method for monitoring and analyzing the operating status of a pantograph, which has the following beneficial effects:

[0057] (1) By synchronously collecting vibration data, contact force data, contact point temperature data, and wear-burn data of the pantograph, the operating status of the pantograph can be comprehensively reflected, improving the accuracy and reliability of monitoring. This multimodal sensor fusion technology can intuitively utilize the relationship between nodes in graph structure data and can be extended to utilize intramodal and intermodal relationships in multimodal problems.

[0058] (2) By synchronously collecting vibration data, contact force data and contact point temperature data through a multimodal sensor system, and combining dynamic calculation to generate dynamic contact force, the changes in contact force can be monitored in real time, thereby improving the response speed and accuracy to abnormal contact force. Furthermore, based on the detection results, the passive early warning strategy can be changed to active control. By analyzing the vibration data and dynamic contact force, the probability of hard point impact can be calculated, and the train speed limit command can be triggered based on the probability value. Through these two intervention control commands, the impact of hard point impact on train operation can be effectively reduced, and the problem of continuous deterioration of pantograph damage due to untimely adjustment can be avoided. This is more conducive to improving train safety and can reduce train downtime caused by malfunctions, thereby improving the overall efficiency of railway transportation.

[0059] (3) By using a multimodal data fusion neural network, vibration data, contact force data, contact point temperature data and wear data are comprehensively analyzed and multimodal fusion feature vectors are output. These vectors are then input into the operation status assessment model for fault diagnosis. This allows for a more comprehensive assessment of the pantograph's operation status. Through intelligent monitoring and prediction technology, the frequency and cost of manual inspection are reduced, and the accuracy of the assessment is improved while avoiding maintenance costs and safety risks caused by the expansion of faults. Attached Figure Description

[0060] Figure 1 This is a schematic diagram illustrating the steps of a pantograph operation status monitoring and analysis method according to the present invention;

[0061] Figure 2 This is a schematic diagram illustrating the determination of whether a pressure control command is triggered in a pantograph operation status monitoring and analysis method according to the present invention.

[0062] Figure 3 This is a schematic diagram illustrating whether an interruption in a pantograph operation status monitoring and analysis method triggers a train speed limit command. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0064] Example 1

[0065] Please see Figure 1-3 This invention provides a method for monitoring and analyzing the operating status of a pantograph, comprising:

[0066] Step 1: Synchronously collect vibration data, contact force data, contact point temperature data, and wear data of the pantograph using a multimodal sensor system.

[0067] It should be noted that the multimodal sensor system includes at least fiber optic grating sensors, airbag pressure sensors, infrared thermal imagers, and laser profilometers, which are used to collect vibration data, contact force data, contact point temperature data, and wear data, respectively.

[0068] Specifically, fiber Bragg grating sensors are embedded in locations such as the pantograph's sliding plate base, tie rod hinge points, and frame support beams. A sampling rate of 1kHz and a wavelength resolution of 1pm can be selected. A demodulation module (such as an SM130 demodulator) converts the wavelength offset into vibration data, and a wavelet denoising algorithm (Daubechies 4 wavelet basis) is used to eliminate high-frequency noise. An airbag pressure sensor is installed on the carbon sliding plate support component of the pantograph to collect contact force data. An infrared thermal imager should be installed in a location where the contact area between the pantograph and the contact wire can be clearly observed to accurately measure the contact point temperature. The optimal installation location needs to be determined based on the specific design and installation space of the pantograph to collect contact point temperature data. A laser profilometer, using a 650nm line laser at a 45° incident angle, scans the sliding plate surface in scanning mode and calculates the wear depth using triangulation principles. ,in, θ is the displacement of the laser point; θ = 45°, φ is the tilt angle of the slide plate, which can be provided by the inertial measurement unit.

[0069] Step 2: Perform time alignment and data preprocessing on the collected data.

[0070] Specifically, the method for time-aligning the collected data is as follows:

[0071] A unified timescale is added to each sensor in the multimodal sensor system, and then the data transmission time difference of each sensor is calculated using an optical fiber transmission delay compensation algorithm. The compensation formula is as follows:

[0072] ;

[0073] Among them, t 补偿 The time difference that needs to be compensated for during data transmission from the sensor; L is the fiber optic length; n 纤 c is the refractive index of the fiber core; t is the speed of light; 处理 The processing delay time of the demodulator during the data transmission process of the sensor;

[0074] By calculating t for each sensor 补偿 Select a reference time point t 参考 Then calculate the time delay for each sensor. t, time delay The formula for calculating t is: t=t 参考 -t 补偿 Based on the time delay of each sensor t adjusts the timestamp of the corresponding sensor to match the reference time point t. 参考 Alignment, if If t is a positive value, the timestamp is adjusted forward. t, if If t is negative, the timestamp will be adjusted backward. By aligning the collected data with time, we can ensure the temporal consistency of different types of data, and ensure that the calculation results at different time points are more accurate and precise.

[0075] Specifically, preprocessing includes at least data normalization preprocessing to eliminate the dimensions of different parameters, which facilitates subsequent calculation and analysis, and may also include data preprocessing forms such as noise reduction.

[0076] Step 3: By comprehensively calculating the processed vibration data, contact force data, and contact point temperature data, the dynamic contact force of the pantograph is generated. The pantograph pressure control command is triggered based on the dynamic contact force. The pantograph pressure control command controls and adjusts the outlet pressure of the pneumatic system, thereby adjusting the pantograph-catenary contact force. When the dynamic contact force is too large, the pressure is reduced; when the dynamic contact force is too small, the pressure is increased.

[0077] Step 4: Analyze the vibration data and dynamic contact force, and calculate the probability of hard point impact; determine whether to trigger the train speed limit command based on the probability of hard point impact.

[0078] Step 5: By comprehensively analyzing wear data, contact point temperature data, and contact force data, the predicted damage depth is obtained; the wear area value of the pantograph friction area is collected through a quantum dot fluorescence monitoring system; and the dynamic weights of dynamic contact force, hard point impact probability, predicted wear depth, and wear area value are calculated using an adaptive model.

[0079] It should be noted that the method for collecting the wear area value of the pantograph friction region using the quantum dot fluorescence monitoring system is as follows: Quantum dots can be uniformly coated on the carbon sliding plate surface in the pantograph friction region beforehand. These quantum dots will detach due to wear during the friction process, thus forming a fluorescence signal distribution map. The area coated with quantum dots is irradiated by an excitation light source, and a fluorescence detector collects the fluorescence signal, converting it into an electrical signal and transmitting it to a signal processor. High-precision three-dimensional point cloud data of the pantograph surface is obtained using a three-dimensional measurement system (such as laser scanning or a digital matrix camera). This data can be combined with the quantum dot fluorescence signal distribution, and the area of ​​the wear region can be calculated through digital modeling and image processing techniques.

[0080] Step 6: Extract features from the preprocessed dynamic contact force, hard point impact probability, predicted wear depth, and wear area value respectively. Then, input the extracted different features and corresponding dynamic weights into the multimodal data fusion neural network to output the multimodal fusion feature vector.

[0081] Step 7: Input the multimodal fusion feature vector into the operation status assessment model, and use the operation status assessment model to assess whether there is an operational fault in the pantograph.

[0082] This invention utilizes a multimodal sensor system to simultaneously collect vibration data, contact force data, and contact point temperature data. Combined with dynamic calculations, it generates dynamic contact force, enabling real-time monitoring of contact force changes. This improves the response speed and accuracy to contact force anomalies and, based on the detection results, transforms a passive early warning strategy into active control. By analyzing vibration data and dynamic contact force, it calculates the probability of hard-point impact and triggers train speed-limiting commands based on the probability value. These two intervention control commands effectively reduce the impact of hard-point impacts on train operation, preventing the continuous deterioration of pantograph damage due to untimely adjustments, thus enhancing train safety and reducing train downtime caused by malfunctions, thereby improving the overall efficiency of railway transportation. Furthermore, by employing a multimodal data fusion neural network, it comprehensively analyzes vibration data, contact force data, contact point temperature data, and wear data, outputting a multimodal fusion feature vector. This vector is then input into an operational status assessment model for fault diagnosis, enabling a more comprehensive evaluation of the pantograph's operational status. Through intelligent monitoring and prediction technologies, it reduces the frequency and cost of manual inspection, improves assessment accuracy, and avoids maintenance costs and safety risks caused by escalating faults.

[0083] Example 2

[0084] In the preferred embodiment of the above-mentioned pantograph operation status monitoring and analysis method, the method for generating the dynamic contact force of the pantograph is as follows:

[0085] The preprocessed vibration data, contact force data, and contact point temperature data are input into the data coupling model for comprehensive calculation, outputting the dynamic contact force of the pantograph. The calculation formula of the data coupling model is as follows:

[0086] ;

[0087] in, F d It is the dynamic contact force of the pantograph; F s It is the static contact force of the pantograph; m iThe vibrational mass of the pantograph in different directions can be obtained through vibration modal analysis or finite element simulation systems. i The index indicates different directions and can be 1, 2, or 3. m 1 represents the component of the pantograph's vibrating mass in the X-axis direction; m 2 represents the component of the pantograph's vibrating mass in the Y-axis direction; m 3 represents the component of the pantograph's vibrating mass in the Z-axis direction; a 1 represents the vibration acceleration component of the pantograph in the X-axis direction; a 2 represents the vibration acceleration component of the pantograph in the Y-axis direction; a 3 represents the vibration acceleration component of the pantograph in the Z-axis direction; β This is the coefficient of thermal expansion, which can be obtained by consulting industry data or calibrating through experiments. Here, we take a value of [value missing]. N / ℃; ρ is the air density, which can be obtained through an air density measurement system. An air density measurement system typically includes components such as a temperature sensor, humidity sensor, pressure sensor, CO2 content sensor, data acquisition system, and display instrument. By accurately measuring parameters such as temperature, humidity, pressure, and CO2 content in the quality environment system, the air density value in the quality measurement environment is determined using the CIPM formula. Here, the value is taken as... kg / m 3 C d This is the aerodynamic drag coefficient, which can be determined through wind tunnel testing; here, it is taken as [value missing]. A represents the frontal area of ​​the vehicle body, which can be obtained by creating a 3D model based on train data and simulating the train's operating environment using modeling software. T 0 is the reference temperature, which can be a normal temperature value, such as 25℃; T c It is the temperature of the pantograph's contact point; v It's vehicle speed, which can be measured by sensors.

[0088] In the formula, m i a i It is used to represent the vibration inertial force term, which can reflect the dynamic characteristics of the pantograph's mechanical mechanism; It is used to represent the thermal expansion force term, which can quantify the effect of thermal expansion on contact force and characterize the temperature-mechanical coupling effect; It is used to represent aerodynamic terms and characterize speed-related air resistance corrections. For example, in the physical field interaction example: vehicle speed increases → aerodynamic drag increases → contact force increases. At the same time, frictional heat generation leads to an increase in the contact point temperature → thermal expansion force is further superimposed. The model provides core algorithm support for the dynamic contact force calculation of high-speed pantographs by characterizing the multi-field coupling of mechanical vibration, thermodynamics, and aerodynamic effects. Through the synergistic effect of multiple data, it can improve the realism and accuracy of the calculated dynamic contact force.

[0089] Furthermore, the method for determining whether the pantograph pressure control command has been triggered is as follows:

[0090] Set the contact force range threshold;

[0091] It should be noted that the contact force range threshold can be trained on historical data using a machine learning model. The contact force range threshold can be dynamically adjusted based on changes in real-time data and environmental conditions, thereby improving the real-time performance of the solution and increasing control precision. It can also be set according to industry standards and specifications.

[0092] Collect dynamic contact forces within a unit time period to form a dynamic contact force sequence. This represents the dynamic contact force at different points in time within a unit time period. The average contact force of multiple dynamic contact forces in the dynamic contact force sequence is calculated, and the average contact force is compared with the contact force interval threshold. If the value of the average contact force does not meet the contact force interval threshold, a pressure control command needs to be triggered. Calculating the average of the dynamic contact force sequence over a shorter unit time period can avoid false triggering caused by occasional data anomalies, thereby ensuring the rationality of monitoring.

[0093] In the above embodiments, by comprehensively considering multiple influencing factors such as static contact force, vibration, thermal expansion, and aerodynamics, the calculation results are more comprehensive and accurate. The dynamic contact force is calculated, providing a basis for real-time monitoring and adjustment of the pantograph. This further improves the accuracy of the calculation, avoids the combined influence of multiple factors that may be ignored by single-factor calculation methods, reduces calculation errors, and improves the reliability of pantograph operation status monitoring. By setting a contact force range threshold and comparing it with the dynamic contact force, the contact force can be easily controlled, thereby effectively preventing the pantograph from continuously deteriorating due to contact force issues and improving the safety of train operation.

[0094] In the preferred embodiment of the above-mentioned pantograph operation status monitoring and analysis method, the method for calculating the hard-point impact probability by analyzing vibration data and dynamic contact force is as follows:

[0095] Vibration data and dynamic contact force are input into the hard-point impact probability model. The hard-point impact probability model can extract the kurtosis coefficient from the vibration data, as shown in the formula: Where K is the kurtosis coefficient, used to measure the sharpness of vibration data. N σ is the number of input samples; μ is the standard deviation of the pantograph's acceleration, reflecting the dispersion of the vibration data; μ is the mean of the pantograph's vibration acceleration components; the kurtosis coefficient is highly sensitive to extreme values ​​in the data and can effectively capture vibration anomalies caused by hard point impacts. A larger kurtosis coefficient indicates that the data distribution has more extreme values, meaning that the probability of hard point impacts is higher. Traditional methods may be affected by the installation location and accuracy of the contact force sensor, and are only applicable to specific contact network structures. In contrast, the kurtosis coefficient method, based on vibration data, is less dependent on the sensor installation location and is applicable to various contact network structures and operating conditions, offering higher adaptability.

[0096] The hard-point impact probability model can extract the contact force fluctuation rate from the dynamic contact force sequence, as shown in the formula: ;in, δ F For contact force fluctuation rate, F d,max and F d,min These are the maximum and minimum values ​​of the dynamic contact force in the dynamic contact force sequence, respectively. F d,avg It is the mean of the dynamic contact forces in the dynamic contact force sequence; then, the hard point impact probability is calculated and output using the determination formula, which is:

[0097] ;

[0098] The method for determining whether a train speed limit command is triggered based on the probability of hard point impact is as follows:

[0099] Set hard point impact assessment thresholds;

[0100] It should be noted that the hard point impact assessment threshold can be trained on historical data using a machine learning model and dynamically adjusted according to changes in real-time data and environmental conditions, thereby improving the real-time performance of the solution and making the control precision higher. It can also be set according to industry standards and specifications.

[0101] When the probability of a hard impact is greater than or equal to the hard impact assessment threshold, a train speed limit command is triggered to reduce the train speed and avoid excessive damage to the pantograph.

[0102] In the above embodiments, combining the kurtosis coefficient and dynamic contact force fluctuation rate enables more accurate identification of hard point impacts. The kurtosis coefficient reflects the distribution pattern of the contact force sequence, while the dynamic contact force fluctuation rate reflects the degree of fluctuation in the contact force. The combination of the two allows for a more comprehensive assessment of the likelihood of hard point impacts; the calculation formula for the probability of hard point impacts is dynamically adjusted according to different ranges of the kurtosis coefficient and dynamic contact force fluctuation rate. This effectively reduces false alarms and missed alarms. Using different calculation formulas under different conditions allows for more accurate identification of hard point impacts, avoiding misjudgments caused by a single indicator. This dynamic adjustment can adapt to different operating conditions and contact force changes, improving the flexibility and accuracy of monitoring.

[0103] Example 3

[0104] In the preferred embodiment of the above-mentioned pantograph operation status monitoring and analysis method, the method for predicting damage depth is as follows:

[0105] The network structure of the wear depth prediction model includes an input layer, an LSTM layer, and an output layer. The input layer...

[0106] Wear data, contact point temperature data, and contact force data from a past unit of time period are input into the input layer of the wear depth prediction model;

[0107] The analysis and calculation are performed using an LSTM layer, and the formula is as follows:

[0108] ,in, d s It is the predicted wear depth of the pantograph. d w It is the wear depth of the pantograph, t 当前 It is the current time point, t 单 It is a unit of time period;

[0109] The output layer outputs the predicted wear depth for a future unit of time period.

[0110] It should be noted that, This indicates the wear depth value of the pantograph at different time points within a previous historical time unit, up to the current time. This represents the dynamic contact force generated at different points in time within a previous historical time unit, up to the current time. This represents the contact point temperature at different points in time within a previous historical time unit, up to the current time. The time unit can be selected as 60 seconds, and the output layer can output the predicted wear depth at multiple points in the future time unit.

[0111] In the above embodiments, the output layer can output the predicted wear depth at multiple time points in the future unit time period. This allows the model to provide more detailed prediction results, which helps to monitor and manage the pantograph's operating status more precisely. It solves the problem that traditional wear depth prediction methods are difficult to capture long-term dependencies and complex nonlinear relationships, resulting in low prediction accuracy. The LSTM layer can effectively integrate information from multiple factors such as wear depth, dynamic contact force, and contact point temperature, providing more comprehensive prediction results. This effectively solves these problems, provides more accurate prediction results, helps to achieve more refined real-time monitoring and management, and improves the reliability and safety of the system.

[0112] Example 4

[0113] In the preferred embodiment of the above pantograph operation status monitoring and analysis method, the method for calculating dynamic weights using an adaptive model is as follows:

[0114] The dynamic contact force, hard point impact probability, predicted wear depth, and wear area values ​​for consecutive time periods in historical data are integrated into a multimodal dataset. This multimodal dataset, along with vehicle speed, wind speed, and contact point temperature at corresponding time points, is then input into an adaptive model to calculate the dynamic weight of each parameter in the multimodal dataset, as shown in the following formula:

[0115] ;

[0116] Where, q z It represents the dynamic weights of different data in the multimodal dataset, where z represents the index of each data point in the multimodal dataset. The index can be 1, 2, 3, or 4, corresponding to the indexes of dynamic contact force, hard point impact probability, predicted wear depth, and wear area value, respectively. For example... q 1 corresponds to the dynamic weight of the dynamic contact force; q 2. Dynamic weights corresponding to the hardpoint impact probability; q 3. Dynamic weights corresponding to the predicted wear depth; q 4. Dynamic weights corresponding to the wear area; S z This represents the parameter sensitivity of different data in a multimodal dataset; for example... S 1 corresponds to the sensitivity of parameters related to dynamic contact force; S 2. Parameter sensitivity corresponding to hard point impact probability; S 3. Sensitivity of parameters corresponding to predicted wear depth; S 4. Parameter sensitivity corresponding to the wear area; E z represents the information entropy of different data in a multimodal dataset; such as E 1 corresponds to the information entropy of dynamic contact force; E 2 corresponds to the information entropy of the hard point impact probability; E3. Information entropy corresponding to the predicted wear depth; E 4. Information entropy corresponding to the wear area.

[0117] It should be noted that parameter sensitivity represents the degree of influence of different data in a multimodal dataset on the pantograph's operating status. This can be determined through industry standards or calculated using experimental data. For example, a series of experiments can be designed to change the values ​​of various parameters (such as dynamic contact force, hard-point impact probability, predicted wear depth, and wear area) while keeping other parameters constant. Data on the pantograph's operating status can be collected under different parameter settings, and the impact of these parameter changes on the pantograph's operating status can be analyzed to calculate the sensitivity of each parameter. Sensitivity can be calculated using methods such as regression analysis and sensitivity analysis. Information entropy represents the information uncertainty of different data in a multimodal dataset. It is calculated by statistically analyzing the distribution of each parameter in historical data and applying the entropy formula from information theory. The formula is:

[0118] ;

[0119] Among them, P z This represents the probability of different parameters in different value ranges within a multimodal dataset. 'u' is the index of the value range, which is a positive integer. This can be achieved by dividing historical data into several value ranges according to requirements and numbering each range using the subscript 'u'. For example, P... 1，u The probability of the dynamic contact force taking the uth value range.

[0120] Furthermore, the extracted different features and their corresponding dynamic weights are input into a multimodal data fusion neural network, which outputs a multimodal fusion feature vector. ,in, It is the wear area value of the wear zone.

[0121] It should be noted that features of different parameters can be extracted using methods known in the field, such as mathematical models like Fourier transform.

[0122] Furthermore, the method for predicting whether the pantograph has a risk of operational failure within a future unit of operating time is as follows:

[0123] Several training samples are constructed using multimodal fusion feature vectors Rr from historical records and their corresponding fault category labels Gl, forming a training sample set. ,in, Rr n Indicates the first n The multimodal fusion feature vector of each training sample; Gl n Indicates the first nFault category labels for each training sample; n This is the index of the training sample in the training sample set, and its value is a positive integer;

[0124] It should be noted that the fault category label Gl can be obtained by first determining the fault type and then calculating the multimodal fusion feature vector Rr corresponding to the fault type, so that the two have a clear correspondence and become a set of training samples.

[0125] The support vector machine model is trained using a training sample set. After training, the multimodal fusion feature vector Rr of new samples is input into the support vector machine model to calculate the classification function of the new samples. f(x) By classification function f （x） The numerical values ​​are used to match the fault type of new samples; classification function f(x) The formula is:

[0126] ;

[0127] Where, ax j The Lagrange multipliers are parameters obtained during SVM training, representing the degree of influence of each training sample on the decision boundary; Rr j Let Gl be the multimodal fusion feature vector of the j-th sample. j Let j be the fault category label for the j-th sample. is the kernel function; b is the bias term, which is a parameter obtained during SVM training and is used to adjust the position of the decision boundary.

[0128] The formula used to calculate the similarity between two samples in the feature space is as follows:

[0129] ;

[0130] Where γ is the kernel function parameter, which controls the width of the kernel function.

[0131] It should be noted that the support vector machine model distinguishes samples and then inputs a new sample's multimodal fusion feature vector. The support vector machine model can compare this new sample with the multimodal fusion feature vector of the samples during training, match the closest sample, find the corresponding fault category label, and thus determine the operational fault.

[0132] In the above embodiments, the use of multimodal fusion feature vectors can integrate data from multiple sensors for fault classification, improving the accuracy and robustness of the classification. It also enables the mapping of nonlinearly separable data to a high-dimensional feature space for classification, thus enhancing the model's classification capability.

[0133] Example 5

[0134] The present invention also discloses a pantograph operation status monitoring and analysis system for implementing the above-mentioned pantograph operation status monitoring and analysis method.

[0135] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0136] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0137] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for monitoring and analyzing the operating status of a pantograph, characterized in that, include: Step 1: Synchronously collect vibration data, contact force data, contact point temperature data, and wear data of the pantograph using a multimodal sensor system; Step 2: Perform time alignment and data preprocessing on the collected data; Step 3: By comprehensively calculating the processed vibration data, contact force data, and contact point temperature data, the dynamic contact force of the pantograph is generated, and the pantograph pressure control command is triggered based on the dynamic contact force. Step 4: Analyze vibration data and dynamic contact force, and calculate the probability of hard point impact; determine whether to trigger a train speed limit command based on the hard point impact probability; the method for calculating the hard point impact probability is as follows: Vibration data and dynamic contact force are input into the hard-point impact probability model. The hard-point impact probability model can extract the kurtosis coefficient from the vibration data, as shown in the formula: Where K is the kurtosis coefficient. N σ is the number of input samples; μ is the standard deviation of the pantograph's acceleration; and μ is the mean of the pantograph's vibration acceleration components. i The number represents the direction, and can be 1, 2, or 3. a 1 represents the vibration acceleration component of the pantograph in the X-axis direction; a 2 represents the vibration acceleration component of the pantograph in the Y-axis direction; a 3 represents the vibration acceleration component of the pantograph in the Z-axis direction; The hard-point impact probability model can extract the contact force fluctuation rate from the dynamic contact force sequence, as shown in the formula: ;in, δ F For contact force fluctuation rate, F d,max and F d,min These are the maximum and minimum values ​​of the dynamic contact force in the dynamic contact force sequence, respectively. F d,avg It is the mean of the dynamic contact forces in the dynamic contact force sequence; then, the hard point impact probability is calculated and output using the determination formula, which is: ； The method for determining whether a train speed limit command is triggered based on the probability of hard point impact is as follows: Set hard point impact assessment thresholds; When the probability of hard impact is greater than or equal to the hard impact assessment threshold, the train speed limit command is triggered. Step 5: By comprehensively analyzing wear data, contact point temperature data, and contact force data, the predicted damage depth is obtained; the wear area value of the pantograph friction area is collected using a quantum dot fluorescence monitoring system; the dynamic weights of dynamic contact force, hard point impact probability, predicted wear depth, and wear area value are calculated using an adaptive model; the method for collecting the wear area value of the pantograph friction area is as follows: quantum dots are uniformly coated on the carbon sliding plate surface in the pantograph friction area beforehand. These quantum dots will fall off due to wear during the friction process, thus forming a fluorescence signal distribution map. The area coated with quantum dots is irradiated by an excitation light source, and the fluorescence detector collects the fluorescence signal and converts it into an electrical signal, which is then transmitted to the signal processor; three-dimensional point cloud data of the pantograph surface is obtained using a three-dimensional measurement system. This data is combined with the quantum dot fluorescence signal distribution, and the area of ​​the wear area is calculated using digital modeling and image processing techniques; Step 6: Extract features from the preprocessed dynamic contact force, hard point impact probability, predicted wear depth, and wear area value respectively. Then, input the extracted different features and corresponding dynamic weights into the multimodal data fusion neural network to output the multimodal fusion feature vector. Step 7: Input the multimodal fusion feature vector into the operation status assessment model, and use the operation status assessment model to assess whether there is an operational fault in the pantograph.

2. The pantograph operation status monitoring and analysis method according to claim 1, characterized in that, The method for time-aligning the collected data is as follows: A unified timescale is added to each sensor in the multimodal sensor system, and then the data transmission time difference of each sensor is calculated using an optical fiber transmission delay compensation algorithm. The compensation formula is as follows: ； Among them, t 补偿 The time difference that needs to be compensated for during data transmission from the sensor; L is the fiber optic length; n 纤 c is the refractive index of the fiber core; t is the speed of light; 处理 The processing delay time of the demodulator during the data transmission process of the sensor; Calculate t for each sensor 补偿 Select a reference time point t 参考 Then calculate the time delay for each sensor. t, time delay The formula for calculating t is: t=t 参考 -t 补偿 ; Based on the time delay of each sensor t adjusts the timestamp of the corresponding sensor to match the reference time point t. 参考 Alignment.

3. The pantograph operation status monitoring and analysis method according to claim 1, characterized in that, The method for generating the dynamic contact force of the pantograph is as follows: The preprocessed vibration data, contact force data, and contact point temperature data are input into the data coupling model for comprehensive calculation, outputting the dynamic contact force of the pantograph. The calculation formula of the data coupling model is as follows: ； in, F d It is the dynamic contact force of the pantograph; F s It is the static contact force of the pantograph; m i It is the vibrating mass of the pantograph in different directions. i The index indicates different directions and can be 1, 2, or 3. m 1 represents the component of the pantograph's vibrating mass in the X-axis direction; m 2 represents the component of the pantograph's vibrating mass in the Y-axis direction; m 3 represents the component of the pantograph's vibrating mass in the Z-axis direction; β ρ is the coefficient of thermal expansion; C is the density of air. d A is the aerodynamic drag coefficient; A is the frontal area of ​​the vehicle body. T 0 is the reference temperature; T c It is the temperature of the pantograph's contact point; v It's the vehicle speed.

4. The pantograph operation status monitoring and analysis method according to claim 3, characterized in that, The method for determining whether the pantograph pressure control command has been triggered is as follows: Set the contact force range threshold; The system collects dynamic contact forces within a unit time period to form a dynamic contact force sequence. It calculates the average contact force of multiple dynamic contact forces in the sequence and compares the average contact force with a contact force interval threshold. If the average contact force does not meet the contact force interval threshold, a pressure control command needs to be triggered.

5. The pantograph operation status monitoring and analysis method according to claim 4, characterized in that, The method for predicting damage depth is as follows: The network structure of the wear depth prediction model includes an input layer, an LSTM layer, and an output layer. The input layer... Wear data, contact point temperature data, and contact force data from a past unit of time period are input into the input layer of the wear depth prediction model; The analysis and calculation are performed using an LSTM layer, and the formula is as follows: ； in, d s It is the predicted wear depth of the pantograph. d w It is the wear depth of the pantograph, t 当前 It is the current time point, t 单 It is a unit of time period; The output layer outputs the predicted wear depth for a future unit of time period.

6. The pantograph operation status monitoring and analysis method according to claim 5, characterized in that, The method for calculating dynamic weights using an adaptive model is as follows: The dynamic contact force, hard point impact probability, predicted wear depth, and wear area values ​​from historical time periods are integrated into a multimodal dataset. This multimodal dataset, along with vehicle speed, wind speed, and contact point temperature at corresponding time points, is then input into an adaptive model to calculate the dynamic weight of each parameter in the multimodal dataset, as shown in the following formula: ； Where, q z is the dynamic weight of different data in the multimodal dataset, and z represents the sequence number of different data in the multimodal dataset. The sequence number can be 1, 2, 3 or 4, which correspond to the sequence number of dynamic contact force, hard point impact probability, predicted wear depth and wear area value, respectively. S z This indicates the parameter sensitivity of different data in a multimodal dataset; E z represents the information entropy of different data in a multimodal dataset.

7. The pantograph operation status monitoring and analysis method according to claim 6, characterized in that, The extracted features and their corresponding dynamic weights are input into a multimodal data fusion neural network, which outputs a multimodal fusion feature vector. ,in, It is the wear area value of the wear zone.

8. The pantograph operation status monitoring and analysis method according to claim 7, characterized in that, The method for predicting whether there is a risk of pantograph failure in the future unit of operating time is as follows: Several training samples are constructed using multimodal fusion feature vectors Rr from historical records and their corresponding fault category labels Gl, forming a training sample set. ,in, Rr n Indicates the first n The multimodal fusion feature vector of each training sample; Gl n Indicates the first n Fault category labels for each training sample; n This is the index of the training sample in the training sample set, and its value is a positive integer; The support vector machine model is trained using a training sample set. After training, the multimodal fusion feature vector Rr of new samples is input into the support vector machine model to calculate the classification function of the new samples. f(x) By classification function f(x) The numerical values ​​are used to match the fault type of new samples; classification function f(x) The formula is: ； Where, ax j For Lagrange multipliers; Rr j Let Gl be the multimodal fusion feature vector of the j-th sample. j For Rr j Corresponding fault category labels, b is the kernel function; b is the bias term. The calculation formula is: ； Where γ is the kernel function parameter, which controls the width of the kernel function.

9. A pantograph operating status monitoring and analysis system, characterized in that: The pantograph operation status monitoring and analysis method is used to implement any one of claims 1-8 above.

Images